MarkovSim：基于有界上下文的 LLM 智能体框架，面向长程计算仿真
MarkovSim: an LLM Agent Harness with Bounded Context for Long-Horizon Computational Simulation
AI 导读
MarkovSim 是一个面向长程计算仿真的 LLM 智能体框架，核心是把上下文管理内嵌进工作流：工作流记录按预定义 schema 写入记忆系统，并在有限状态机的状态转移边界处构建紧凑表示作为智能体上下文。该框架在有限元分析上的六个基准问题中，18 次运行通过 16 次，能自行纠正求解器操作与模型物理中的错误，仅在需要工程判断时咨询工程师，且全程上下文保持有界。
正文 · 原文
Abstract:Recent studies have applied large language model (LLM) agent systems to computational simulation workflows. Such workflows are long-horizon, so the information generated can exceed what an LLM can reason over at once, its context window. Context management is therefore essential, yet existing systems have either neglected it or treated it as a supporting mechanism rather than a subject of study. This paper presents MarkovSim, an LLM agent harness for long-horizon computational simulation, whose core is a context management scheme built into the workflow. The workflow record is written to a memory system with a predefined schema, and at workflow-defined boundaries, namely the state transitions of a finite-state machine a compact representation of the solution progress is constructed from it as the agent's context. Instantiated on finite element analysis and evaluated on six benchmark problems, MarkovSim passed 16 of 18 runs and corrected its own faults in both the solver operations and the model physics. The engineer was consulted only where engineering judgement was needed. The context remained bounded throughout every run, regardless of its length.
| Comments: | 26 pages,15 figures,not yet submitted to any journal |
| Subjects: | Computational Engineering, Finance, and Science (cs.CE); Artificial Intelligence (cs.AI) |
| MSC classes: | 65-04 |
| ACM classes: | I.2.11; J.6 |
| Cite as: | arXiv:2610.35349 [cs.CE] |
| (or arXiv:2610.35349v1 [cs.CE] for this version) | |
| https://doi.org/10.48550/arXiv.2610.35349 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Wenbo Chen [view email]
[v1]
Mon, 26 Oct 2026 12:41:39 UTC (2,454 KB)
来源：arXiv cs.AI · arxiv.org